[Mlir-commits] [mlir] [mlir][vector] Tighten the semantics of vector.{load|store} (PR #135151)

[Andrzej Warzyński]

================
@@ -718,18 +718,6 @@ func.func @vector_load_i8_with_offset(%arg0 : memref<?x?xi8>) -> vector<[16]x[16
 
 // -----
 
-// CHECK-LABEL: @vector_load_i8_from_rank_1_memref(
-// CHECK-SAME:                                     %[[MEMREF:.*]]: memref<?xi8>)
-// CHECK: %[[C0:.*]] = arith.constant 0 : index
-// CHECK: arm_sme.tile_load %[[MEMREF]][%[[C0]]] : memref<?xi8>, vector<[16]x[16]xi8>
----------------
banach-space wrote:

I need a few more cycles for this, but have posted a draft here: https://github.com/llvm/llvm-project/pull/135396

Let me know if you plan to review this further. If not, I will land this to unblock https://github.com/llvm/llvm-project/pull/134389.

Thanks for the pointer, Ben! 🙏🏻 

https://github.com/llvm/llvm-project/pull/135151